使用 TPU 模式时如何从 Google Colaboratory 保存 Tensorflow 检查点文件？答案

【问题标题】：How to save a Tensorflow Checkpoint file from Google Colaboratory in when using TPU mode?使用 TPU 模式时如何从 Google Colaboratory 保存 Tensorflow 检查点文件？
【发布时间】：2019-03-30 21:19:29
【问题描述】：

当我使用saver = tf.train.Saver() 和save_path = saver.save(session, "checkpointsFolder/checkpoint.ckpt")时

我收到 UnimplementedError (see above for traceback): File system scheme '[local]' not implemented 错误

这是完整的错误

---------------------------------------------------------------------------
UnimplementedError                        Traceback (most recent call last)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/client/session.py in _do_call(self, fn, *args)
   1333     try:
-> 1334       return fn(*args)
   1335     except errors.OpError as e:

/usr/local/lib/python3.6/dist-packages/tensorflow/python/client/session.py in _run_fn(feed_dict, fetch_list, target_list, options, run_metadata)
   1318       return self._call_tf_sessionrun(
-> 1319           options, feed_dict, fetch_list, target_list, run_metadata)
   1320 

/usr/local/lib/python3.6/dist-packages/tensorflow/python/client/session.py in _call_tf_sessionrun(self, options, feed_dict, fetch_list, target_list, run_metadata)
   1406         self._session, options, feed_dict, fetch_list, target_list,
-> 1407         run_metadata)
   1408 

UnimplementedError: File system scheme '[local]' not implemented (file: 'checkpointsBook2Vec5Inputs')
     [[{{node save/SaveV2}} = SaveV2[dtypes=[DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_INT32, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT], _device="/job:tpu_worker/replica:0/task:0/device:CPU:0"](_recv_save/Const_0, save/SaveV2/tensor_names, save/SaveV2/shape_and_slices, embeddings, embeddings/Shampoo, embeddings/Shampoo_1, embeddings/Shampoo_2, epochCount, softmax_biases, softmax_weights, softmax_weights/Shampoo, softmax_weights/Shampoo_1, softmax_weights/Shampoo_2)]]

During handling of the above exception, another exception occurred:

UnimplementedError                        Traceback (most recent call last)
<ipython-input-22-ca87cd5e5739> in <module>()
     48             print('recEpoch_indexA is', recEpoch_indexA)
     49 
---> 50             save_path = saver.save(session, "checkpointsBook2Vec5Inputs/Research2VecCS4.ckpt") #Save checkpoint
     51             print( 'epochCount.eval() is ', epochCount.eval() )
     52 

/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py in save(self, sess, save_path, global_step, latest_filename, meta_graph_suffix, write_meta_graph, write_state, strip_default_attrs)
   1439           model_checkpoint_path = sess.run(
   1440               self.saver_def.save_tensor_name,
-> 1441               {self.saver_def.filename_tensor_name: checkpoint_file})
   1442 
   1443         model_checkpoint_path = compat.as_str(model_checkpoint_path)

/usr/local/lib/python3.6/dist-packages/tensorflow/python/client/session.py in run(self, fetches, feed_dict, options, run_metadata)
    927     try:
    928       result = self._run(None, fetches, feed_dict, options_ptr,
--> 929                          run_metadata_ptr)
    930       if run_metadata:
    931         proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)

/usr/local/lib/python3.6/dist-packages/tensorflow/python/client/session.py in _run(self, handle, fetches, feed_dict, options, run_metadata)
   1150     if final_fetches or final_targets or (handle and feed_dict_tensor):
   1151       results = self._do_run(handle, final_targets, final_fetches,
-> 1152                              feed_dict_tensor, options, run_metadata)
   1153     else:
   1154       results = []

/usr/local/lib/python3.6/dist-packages/tensorflow/python/client/session.py in _do_run(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)
   1326     if handle is None:
   1327       return self._do_call(_run_fn, feeds, fetches, targets, options,
-> 1328                            run_metadata)
   1329     else:
   1330       return self._do_call(_prun_fn, handle, feeds, fetches)

/usr/local/lib/python3.6/dist-packages/tensorflow/python/client/session.py in _do_call(self, fn, *args)
   1346           pass
   1347       message = error_interpolation.interpolate(message, self._graph)
-> 1348       raise type(e)(node_def, op, message)
   1349 
   1350   def _extend_graph(self):

UnimplementedError: File system scheme '[local]' not implemented (file: 'checkpointsBook2Vec5Inputs')
     [[node save/SaveV2 (defined at <ipython-input-15-c14caac2081d>:45)  = SaveV2[dtypes=[DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_INT32, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT], _device="/job:tpu_worker/replica:0/task:0/device:CPU:0"](_recv_save/Const_0, save/SaveV2/tensor_names, save/SaveV2/shape_and_slices, embeddings, embeddings/Shampoo, embeddings/Shampoo_1, embeddings/Shampoo_2, epochCount, softmax_biases, softmax_weights, softmax_weights/Shampoo, softmax_weights/Shampoo_1, softmax_weights/Shampoo_2)]]

Caused by op 'save/SaveV2', defined at:
  File "/usr/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/usr/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py", line 16, in <module>
    app.launch_new_instance()
  File "/usr/local/lib/python3.6/dist-packages/traitlets/config/application.py", line 658, in launch_instance
    app.start()
  File "/usr/local/lib/python3.6/dist-packages/ipykernel/kernelapp.py", line 477, in start
    ioloop.IOLoop.instance().start()
  File "/usr/local/lib/python3.6/dist-packages/zmq/eventloop/ioloop.py", line 177, in start
    super(ZMQIOLoop, self).start()
  File "/usr/local/lib/python3.6/dist-packages/tornado/ioloop.py", line 888, in start
    handler_func(fd_obj, events)
  File "/usr/local/lib/python3.6/dist-packages/tornado/stack_context.py", line 277, in null_wrapper
    return fn(*args, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/zmq/eventloop/zmqstream.py", line 440, in _handle_events
    self._handle_recv()
  File "/usr/local/lib/python3.6/dist-packages/zmq/eventloop/zmqstream.py", line 472, in _handle_recv
    self._run_callback(callback, msg)
  File "/usr/local/lib/python3.6/dist-packages/zmq/eventloop/zmqstream.py", line 414, in _run_callback
    callback(*args, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/tornado/stack_context.py", line 277, in null_wrapper
    return fn(*args, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/ipykernel/kernelbase.py", line 283, in dispatcher
    return self.dispatch_shell(stream, msg)
  File "/usr/local/lib/python3.6/dist-packages/ipykernel/kernelbase.py", line 235, in dispatch_shell
    handler(stream, idents, msg)
  File "/usr/local/lib/python3.6/dist-packages/ipykernel/kernelbase.py", line 399, in execute_request
    user_expressions, allow_stdin)
  File "/usr/local/lib/python3.6/dist-packages/ipykernel/ipkernel.py", line 196, in do_execute
    res = shell.run_cell(code, store_history=store_history, silent=silent)
  File "/usr/local/lib/python3.6/dist-packages/ipykernel/zmqshell.py", line 533, in run_cell
    return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/IPython/core/interactiveshell.py", line 2718, in run_cell
    interactivity=interactivity, compiler=compiler, result=result)
  File "/usr/local/lib/python3.6/dist-packages/IPython/core/interactiveshell.py", line 2822, in run_ast_nodes
    if self.run_code(code, result):
  File "/usr/local/lib/python3.6/dist-packages/IPython/core/interactiveshell.py", line 2882, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-15-c14caac2081d>", line 45, in <module>
    saver = tf.train.Saver()
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", line 1102, in __init__
    self.build()
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", line 1114, in build
    self._build(self._filename, build_save=True, build_restore=True)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", line 1151, in _build
    build_save=build_save, build_restore=build_restore)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", line 792, in _build_internal
    save_tensor = self._AddSaveOps(filename_tensor, saveables)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", line 284, in _AddSaveOps
    save = self.save_op(filename_tensor, saveables)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", line 202, in save_op
    tensors)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/gen_io_ops.py", line 1690, in save_v2
    shape_and_slices=shape_and_slices, tensors=tensors, name=name)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper
    op_def=op_def)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/util/deprecation.py", line 488, in new_func
    return func(*args, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/ops.py", line 3274, in create_op
    op_def=op_def)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/ops.py", line 1770, in __init__
    self._traceback = tf_stack.extract_stack()

UnimplementedError (see above for traceback): File system scheme '[local]' not implemented (file: 'checkpointsBook2Vec5Inputs')
     [[node save/SaveV2 (defined at <ipython-input-15-c14caac2081d>:45)  = SaveV2[dtypes=[DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_INT32, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT], _device="/job:tpu_worker/replica:0/task:0/device:CPU:0"](_recv_save/Const_0, save/SaveV2/tensor_names, save/SaveV2/shape_and_slices, embeddings, embeddings/Shampoo, embeddings/Shampoo_1, embeddings/Shampoo_2, epochCount, softmax_biases, softmax_weights, softmax_weights/Shampoo, softmax_weights/Shampoo_1, softmax_weights/Shampoo_2)]]

查找这个错误，我发现如下：

来自谷歌官方 TPU 调试指南

https://cloud.google.com/tpu/docs/troubleshooting

错误信息

InvalidArgumentError：未实现：文件系统方案“[本地]”不是实施

详情

所有输入文件和模型目录必须使用云存储存储桶路径（gs://bucket-name/...），并且此存储桶必须是可访问的来自 TPU 服务器。注意所有的数据处理和模型检查点在 TPU 服务器上执行，而不是在本地机器上。有关如何正确配置云存储以供使用的信息使用 TPU，请参阅连接到云存储桶的指南。

有类似问题的其他人

TPU local Filesystem doesn't exist?

本地文件系统在 Cloud TPU 上不可用。模型目录（检查点等）和输入数据应存储在 Google Cloud Storage（并以“gs://”为前缀）。

这里有更多细节

https://cloud.google.com/tpu/docs/storage-buckets

但是，我没有 Google Cloud 服务，我只是使用 Google Colab。有没有办法在 TPU 模式下保存 Tensorflow 检查点？

【问题讨论】：

标签： python tensorflow google-colaboratory google-cloud-tpu

【解决方案1】：

另一种方法是使用 Keras 重写模型并将 tf.contrib.tpu.keras_to_tpu_model(..) 与 tf.contrib.tpu.TPUDistributionStrategy(...) 一起使用。这是小代码sn-p：

def get_model():
  return keras.Sequential([
    keras.layers.Dense(10, input_shape=(4,), activation=tf.nn.relu, name = "Dense_1"),
    keras.layers.Dense(10, activation=tf.nn.relu, name = "Dense_2"),
    keras.layers.Dense(3, activation=None, name = "logits"),
    keras.layers.Dense(3, activation=tf.nn.softmax, name = "softmax")
  ])

dnn_model = get_model()

dnn_model.compile(optimizer=tf.train.AdagradOptimizer(learning_rate=0.1), 
              loss='sparse_categorical_crossentropy',
              metrics=['sparse_categorical_crossentropy'])

tpu_model = tf.contrib.tpu.keras_to_tpu_model(
    dnn_model,
    strategy=tf.contrib.tpu.TPUDistributionStrategy(
        tf.contrib.cluster_resolver.TPUClusterResolver(TPU_ADDRESS)))

# Train the model
tpu_model.fit(
  train_x, train_y,
  steps_per_epoch = steps_per_epoch,
  epochs=epochs,
)

tpu_model.save_weights('./saved_weights.h5', overwrite=True)

【讨论】：

这是否也保存了优化器设置？
没有。它只是保存了最终学习到的权重，可以使用以下方法将其加载到同一模型中：prediction_model = get_model() prediction_model.load_weights('./saved_weights.h5')

【解决方案2】：

您可以在free tier 下创建一个Google Cloud 帐户，然后再创建一个GCS bucket。完成此操作后，您可以通过执行以下操作在 Colab 中对自己进行身份验证，以便从 Colab 获得对 GCS 存储桶的写入权限：

from google.colab import auth
auth.authenticate_user()

这是一个使用 Cloud TPU 和 GCS 的 sample Colab notebook。

【讨论】：

看起来是免费试用，12 个月/300 美元。试用结束后是否仍然可以创建/使用存储桶？
是的，但我相信不是免费的。